Seriously? Just a quick Google suggests otherwise:
36,900,000 for indices vs 29,200,000 for indexes
1,440,000,000 for data vs 11,900,000 for datums (Heck, many people aren't familiar with "datum" to begin with...)
7,710,000 for corpora vs 239,000 for corpuses
Linguistics student or not, do you have anything more than "people you talk to use these unusual forms" to base all this on?